DE69737683T2 - ANTIBODIES TO THE ED-B DOMAIN OF FIBRONECTIN, THEIR PREPARATION AND USES - Google Patents

Abstract

According to the present invention there is provided a specific binding member which is specific for and binds directly to the ED-B oncofoetal domain of fibronectin (FN). The invention also provides materials and methods for the production of such binding members.

Description

Background of the invention

These The invention relates to specific binding elements for a fetal isoform of fibronectin, ED-B, which is also found in the developing neovasculature of tumors as expressed by both immunocytochemistry and also shown by in vivo tumor targeting. It concerns also materials and procedures that are specific to such Relate binding elements.

The primary The goal of most existing forms of tumor therapy is, if possible kill many cells forming the tumor. The limited success, which has been achieved with chemo and radiotherapy is on the relative lack of specificity treatment and the tendency for toxic side effects to normal Attributed to tissue. One Way, as the tumor selectivity The therapy can be improved, it is the means to the tumor over Binding protein for usually a binding domain an antibody includes, with specificity for a Marker antigen expressed on the surface of the tumor, but missing in normal cells, feed. This form of directed Therapy, which is casually called "magic balls", is mainly based on monoclonal antibodies (MAKs) has been illustrated by rodents that are called for tumor-associated antigens that express on the cell surface become. Such MAKs can either chemically to the cytotoxic moiety (e.g., a toxin or a drug) or may be conjugated to a recombinant fusion protein in which the genes coding for the MAK and the Toxin are bound together and are expressed one behind the other.

Of the Approach of the "magic balls" has a limited, if also significant effect in the treatment of human cancer especially when targeting tumors of lymphoid origin, where most malignant cells are for the therapeutic dose in the bloodstream are freely available. However, the treatment of solid tumors remains a serious one clinical problem, as just a tiny fraction of the total cell mass, especially the cells in the extreme Edge of the tumor therapeutic immunoconjugates in the bloodstream is exposed, these peripheral targets a so-called "binding site barrier" to the tumor interior (Juweid et al., Cancer Res. 52, 5144-5135 (1992)). Within The tumor is the tissue architecture in fibrous flow and tightly packed Tumor cells generally too dense to allow the penetration of molecules in the size range of antibodies to enable. Furthermore, it is known that tumors due to the lack of lymphatic drainage an elevated one have interstitial pressure, which is also the influx of exogenous molecules prevented. For a youngest Report on the factors influencing the uptake of therapeutic agents in Tumors affect, see R. Jain, Sci. At the. 271, 58-65 (1994).

Even though it is aimed at treating solid tumors by targeting tumor-associated antigens are obvious limits, they share Tumors a property that is an alternative antigen target for antibody therapy offers. Once they have exceeded a certain size, tumors hang generally from an independent Blood supply decreases, so that sufficient oxygen and nutrients can sustain growth. When this blood supply is compromised or occluded, there is a realistic potential in this Process starving thousands of tumor cells. If there is a tumor developed, it turns into an angiogenic phenotype, creating a diverse scheme is produced by angiogenic factors acting on adjacent capillary endothelial cells act, causing them to proliferate and migrate. The structure This newly formed blood vessel is highly disorganized with dead ends and cave formations, the increased permeability to lead, in marked contrast to the ordered structure of capillaries in normal tissue. Induction of angiogenesis is of an up-regulation accompanied by the expression of certain cell surface antigens, from many also occur in the vasculature of normal tissues. The Identification of antigens responsible for the neovasculature of tumors are unique, is the most important limiting factor in the development a generic treatment of solid tumors by targeting vessels. The Antigen, which is the subject of the present invention, is concerned directly with this problem.

In the course of tumor progression, the extracellular matrix of the surrounding tissue is remodeled by two major processes: (1) the proteolytic degradation of extracellular matrix components of normal tissue, and (2) the de novo synthesis of extracellular matrix components by both tumor cells and stromal cells be activated by tumor-induced cytokines. These two processes generate an "extracellular tumor matrix" in the steady state which provides a more suitable environment for tumor progression and differs qualitatively and quantitatively from those of normal tissues Among the components of this matrix are the large isoforms of tenascin and fibronectin (FN), the beschrei Use of these proteins as isoforms recognizes their extensive structural heterogeneity, which is brought about at the transcriptional, posttranscriptional and posttranslational level (see below). One of the isoforms of fibronectin, the so-called B + isoform (B-FN), is the subject of the present invention.

fibronectins (FN) are high molecular weight multifunctional glycoprotein components both extracellular Matrix as well as from body fluids. They are involved in many biological processes, e.g. the Creation and maintenance of normal cell morphology, cell migration, hemostasis and thrombosis, wound healing and oncogenic transformation (reports see Alitalo et al. (1982), Yamada (1983), Hynes (1985), Ruoslahti et al. (1988), Hynes (1990); Owens et al. (1986)). Structural diversity at FN becomes by alternative splicing three regions (ED-A, ED-B and IIICS) of the primary FN transcript (Hynes (1985), Zardi et al. (1987)), with at least 20 different Isoforms are produced, some of them in tumor and normal Tissues are expressed differently. As is the case with tissue and development-specific manner is known, that the splice pattern of FN pre-mRNA in transformed cells and deregulated in malignancies (Castellani et al. (1986); Borsi et al. (1987); Vartio et al., (1987); Zardi et al. (1987); Barone et al. (1989); Carnemolla et al. (1989), Oyama et al. (1989) (1990); Borsi et al. (1992b)). Actually the FN isoforms containing ED-A, ED-B and IIICS sequences, in transformed and malignant tumor cells expressed to a greater extent than in normal cells. In particular, the FN isoforms which containing the ED-B sequence (B + isoform), most strongly expressed in fetal and tumor tissues, as well as during wound healing, but its expression in normal tissue is limited (Norton et al. (1987); Schwarzbauer et al. (1987); Guturan and Kornblihtt (1987), Carnemolla et al. (1989); Ffrench-Constant et al. (1989) Ffrench-Constant and Hynes (1989), Laitinen et al. (1991)). B + -FN molecules are not in mature containers but are normal in angiogenic blood vessels (e.g., development of the endometrium), pathological (e.g., in diabetic retinopathy) and upregulated tumor development (Castellani et al., 1994).

The ED-B sequence is complete Type III homology repeat, for which encodes a single exon and which comprises 91 amino acids. In contrast to the alternatively spliced IIICS isoform, which is a cell-type specific Binding site includes, is the biological function of A + - and The B + isoform still gives cause for speculation (Humphries et al. (1986)).

The presence of the B + isoforms themselves constitutes a tumor-induced neoantigen, but in addition, ED-B expression exposes a normally cryptic antigen within type III repeat 7 (before ED-B); since this epitope is not exposed in FN molecules lacking ED-B, it follows that ED-B expression induces the expression of neoantigens both directly and indirectly. This cryptic antigenic site constitutes the target of a monoclonal antibody (MAb) called BC-1 (Carnemolla et al., 1992). The specificity and biological properties of this MAK are in EP 0,344,134 B1 and the same is obtainable from the hybridoma available from the European Collection of Animal Cell Cultures, Porton Down, Salisbury, UK under the number 88042101. The MAK has been used successfully to localize the angiogenic blood vessels of tumors without cross-reactivity with mature vascular endothelium, illustrating the potential of FN isoforms for vascular targeting using antibodies.

Indeed certain reservations remain of specificity of BC-1 MAK upright. The fact that BC-1 is not directly the B + isoform has raised the question of whether or not in some tissues the epitope recognized by BC-1 without the presence of ED-B can be unmasked and therefore indirectly to unwanted cross-reactivity of BC-1 leads. Furthermore, BC-1 is strictly specific to the human B + isoform, which means that animal studies on biodistribution and tumor localization from BC-1 not possible are. Although polyclonal antibodies against recombinant fusion proteins containing the B + isoform, (Peters et al. (1995)), they only react with FN treated with N-glycanase to the epitope unmask.

One Another common problem with the use of monoclonal Is mouse antibodies the response of the human anti-mouse antibody (RAMA) (Schroff et al. (1985), Dejager et al. (1988)). HAMA answers have a number of effects, of neutralization of the administered antibody, what leads to a reduced therapeutic dose, over allergic responses, Serum disease, including kidney damage.

Although polyclonal antisera that reacts with recombinant ED-B have been identified (see above), isolation of MAbs with the same specificity as BC-1 has generally proved difficult after immunization of mice because human and mouse ED-B Proteins show virtually 100% sequence homology. The human protein can therefore look like a self-antigen to mice, which then no immune response to it. Indeed, in over ten years of intensive research in this area, only a single MAb has been identified with indirect reactivity to the B + -FN isoform (BC-1), with none recognizing ED-B directly. It is almost certainly significant that the specificity of BC-1 is for a cryptic epitope exposed as a consequence of ED-B, and not for a portion of ED-B itself that is likely to be absent in mouse FN and therefore of Immune system of the mouse is not recognized as "own".

Implementation of the present invention has been made using an alternative strategy to the previously used strategies, and where no prior immunization with fibronectin or ED-B is required, antibodies specific for the ED-B isoform are considered to be single-chain Fvs (ScFvs) have been obtained from libraries of human variable variable regions presented on the surface of filamentous bacteriophages (Nissim et al. (1994), see also WO 92/01047 . WO 92/20791 . WO 93/06213 . WO 93/11236 . WO 93/19172 ).

The Inventors have found using an antibody phage library, that specific scFvs by both direct selection on recombinant FN fragments containing the ED-B domain as well recombinant ED-B itself can be isolated when these antigens on a solid surface stratified ("panning"). The same sources Antigens have also been used successfully to create a "second Generation "from scFvs with improved properties compared to the parent clones in one Process of "affinity maturity" to produce. The inventors have found that the isolated scFvs without previous treatment with N-glycanase strong and specific with the B + isoform of human FN react.

In Antitumor applications have the antigen binding domains of the human antibody, provided by the present invention, the advantage not trigger to be a HAMA answer. Further, they are as illustrated herein is used for immunohistochemical analysis of tumor tissue, both in vitro as well as in vivo. These and other applications are discussed in detail herein and are for the average Professional obviously.

Zang et al., Matrix Biology, Vol. 14, 623-633 (1994) describe one polyclonal antibodies, the one against an extra domain B of fibronectin bred from dogs has been.

JP02076598 and JP 04169195 describe monoclonal antibodies to fibronectin.

terminology

Specific binding (pair) element

This describes one element of a pair of molecules, the binding specificity for each other to have. The elements of a specific binding pair can be more natural His descent is complete or complete or partially synthetically produced. An element of the couple of molecules has an area on its surface or a cave that specific to a specific spatial and polar organization the other element of the pair of molecules binds and is therefore complementary to it. Therefore the elements of the pair have the property of being specific to each other to bind. examples for such forms of specific binding pairs are antigen-antibody, biotin-avidin, Hormone Hormone Receptor, Receptor Ligand, Enzyme Substrate.

antibody

This describes an immunoglobulin, whether natural or partially or completely synthetic produced. The term also covers any polypeptide or Protein off that over a binding domain features, which is an antibody binding domain or homologous to it. these can of natural Sources derived or partially or completely synthetically produced become. examples for antibody are the immunoglobulin isotypes and their isotype subclasses, fragments, the one antigen binding domain include, e.g. Fab, scFv, Fv, dAb, Fd, and diabody.

It is possible to use monoclonal and other antibodies and to use recombinant DNA technology techniques to produce other antibodies or chimeric molecules that retain the specificity of the original antibody. Such methods may include introducing DNA encoding the variable immunoglobulin region, or complementarity determining regions (CDRs), of one antibody into the constant regions or constant regions plus framework regions of another immunogen include bulin. See eg EP-A-184187 . UK-A-2188638 or EP-A-239,400 , A hybridoma or other cell that produces an antibody may or may not undergo gene mutation or other alterations that may alter the binding specificity of antibodies produced.

Since antibodies can be modified in a variety of ways, the term "antibody" should be delineated to cover any specific binding element or substance that has a binding domain with the required specificity, therefore, this term covers antibody fragments, derivatives, functional equivalents and homologues of antibodies comprising any polypeptide comprising an immunoglobulin binding domain, whether natural or wholly or partially synthetic, and therefore also containing chimeric molecules comprising an immunoglobulin binding domain, or equivalent thereof, fused to another polypeptide The cloning and expression of chimeric antibodies is described in EP-A-0210694 and EP-A-0125023 described.

It has been shown that fragments of an entire antibody can fulfill the function of binding antigens. Examples of binding fragments are (i) the Fab fragment consisting of VL, VH, CL and CH1 domains; (ii) the Fd fragment consisting of VH and CH1 domains; (iii) the Fv fragment consisting of VL and VN domains of a single antibody; (iv) the dAb fragment (Ward et al. (1989)) consisting of a VN domain; (v) isolated CDR regions; (vi) F (ab ') ₂ fragments, bivalent fragments comprising two linked Fab fragments; (vii) single chain Fv molecules (scFv) wherein a VN domain and a VL domain are linked via a peptide linker allowing the two domains to associate to form an antigen binding site (Bird et al. (1988) , Huston et al. (1988)); (viii) bispecific scFv dimers ( PCT / US92 / 09965 ); and (ix) "diabodies", multivalent or multispecific fragments generated by gene fusion ( WO 94/13804 ; Holliger et al. (1993)).

Diabodies are multimers of polypeptides, each polypeptide comprising a first domain comprising an immunoglobulin light chain binding region and a second domain containing an immunoglobulin heavy chain binding region, the two domains being linked (eg via a peptide linker), but are unable to associate with each other to form an antigen binding site; Antigen binding sites are formed by associating the first domain of a polypeptide within each multimer with the second domain of another polypeptide within the multimer ( WO 94/13804 ).

If bispecific antibodies can be used this conventional bispecific antibodies in a number of ways (Holliger and Winter (1993)), e.g. chemically or from hybrid hybridomas, or can be prepared may be any of the bispecific antibody fragments which previously mentioned. It may be preferred to use scFv dimers or diabodies instead of whole antibodies to use. diabodies and scFvs can without FC region, using only variable domains which are potentially the effects of anti-idiotypic reactions to reduce. Other forms of bispecific antibodies include the single-chain "Janusine" available from Traunecker et al. (1991).

Ice-specific diabodies, in contrast to bispecific whole antibodies, may also be particularly useful because they are easy to prepare and can be expressed in E. coli. Diabodies (and many other polypeptides, such as antibody fragments) having suitable binding specificities can be conveniently prepared using phage display ( WO 94/13804 ) are selected from libraries. If one arm of the diabody is to be kept constant, eg with a specificity directed against antigen X, then a library in which the other arm is variable can be generated and an antibody of suitable specificity can be selected.

Antigen binding domain

This describes the part of an antibody which encompasses the region specific to the entire antigen or part of it binds and is complementary to it. If an antigen is great can be an antibody bind only to a certain part of the antigen, its part epitope is called. An antigen binding domain may be replaced by one or more provided variable antibody domains become.

Preferably an antigen binding site comprises a variable antibody light chain (VL) region and a variable antibody heavy chain (VH) region.

Specific

This denotes the situation in which an element of a specific Binding pair shows no significant binding to molecules, who are not their specific binding partner. The term is even if, for example, an antigen binding domain for a particular Epitope is specific, which is carried by a number of antigens is, in which case the specific binding element, the the antigen binding domain wearing, will be able to bind to the various antigens, who carry the epitope.

Functionally equivalent variant form

This denotes a molecule (the variant) that, despite being structural differences to one another molecule (the parent molecule) has some significant homology and also at least has some biological function of the parent molecule, e.g. the ability, to bind a specific antigen or epitope. Variants can be found in Form fragments, derivatives or mutants. A variant, a derivative or a mutant can be obtained by modification of the parent molecule Addition, deletion, substitution or insertion of one or more amino acids or by binding another molecule. These changes can be performed at the nucleotide or protein level. For example For example, the encoded polypeptide may be a Fab fragment which is then attached a Fc tail is bound from another source. Alternatively For example, a marker, such as e.g. an enzyme, fluorescein etc., tethered become.

Summary of the present invention

According to the present The invention as claimed in claim 1 is a specific one Binding element provided that is specific for the oncofetal domain of fibronectin (FN) is and binds directly to them.

specific Binding elements according to the invention bind directly to the ED-B domain. In one embodiment binds a specific binding element after treatment of the FN the protease thermolysin to one, some or all FN containing ED-B. In a further embodiment binds a specific binding element to one, some or all FN-containing type III homology repeats containing the ED-B domain include. Known FNs are described in two reports by Carnemolla et al. (1989) (1992). References to "all FNs that contain ED-B" can be considered as Reference to all FNs listed in these reports as FNs containing ED-B are identified.

The specific binding element preferably binds human ED-B and preferably B + FN of at least one other species, e.g. Mouse, rat and / or chicken. Preferably, the specific binding pair element is capable of both human fibronectin-ED-B and a non-human To bind fibronectin-ED-B, e.g. that of a mouse, so that testing and to allow analysis of the sbp element in an animal experiment.

specific Binding pair elements according to the present Invention bind fibronectin-ED-B without being deposited with the freely available Antibody BC-1 to compete, which is discussed elsewhere herein. BC-1 is for a human B + isoform strictly specific. Specific binding pair elements according to the present Invention do not bind the same epitope as BC-1.

The Binding of a specific binding element according to the present invention B + FN may be through the ED-B domain be inhibited.

The binding domain, when measured as a purified monomer, has a dissociation constant (Kd) of 6 x 10 ^-8 M or less for ED-B-FN.

In In one aspect of the present invention, the binding domain reacts with fibronectin-ED-B without prior treatment of fibronectin-ED-B with N-glycanase, i.e. she is able to tie this.

Specific binding pair members of the present invention can be provided as isolates or in purified form, ie, in a preparation or formulation that is free of other specific binding pair members, eg, antibodies or antibody fragments, or free from other specific binding pair members capable of To specifically bind fibronectin-ED-B. Preferably, the specific binding members of the present invention will be in substantially pure form Semi asked. They may be "monoclonal" in the sense of being derived from a single clone and are not limited to antibodies obtained using traditional hybridoma technology As discussed, specific binding pair members of the present invention may be constructed using bacteriophage display methods and / or expression in recombinant, eg bacterial, host cells are obtained.

The specific binding element comprises an antibody. The specific binding member may comprise a polypeptide sequence in the form of an antibody fragment, such as a single-chain Fv (scFv). Other forms of antibody fragments may also be used, such as Fab, Fab ', F (ab') ₂ , Facb, Facb, or a diabody (Winter and Milstein (1991); WO 94/13804 ). The specific binding member may be in the form of whole antibody, the whole antibody may be in any of the antibody isotype forms, eg, IgG, IgA, IgD, IgE, and IgM, and any of the isotype subclass forms, eg, IgG1 or IgG4.

Of the antibody may be of any origin, e.g. human, murine, ovines or lapine origin. Other lineages are known to professionals. Preferably, the antibody is of human origin. By "human" one understands one Antibody, the partial or complete from a human cDNA, protein or peptide library. This term includes humanized peptides and nonhuman proteins Of origin, which have been modified to human the antibody molecule To lend properties and thus enable the molecule, the defense mechanisms to bypass the human immune system.

The specific binding member may also be in the form of a genetically engineered antibody, eg a bispecific antibody molecule (or a fragment such as F (ab ') ₂ ) having an antigen binding arm (ie, a specific domain) against fibronectin ED-B as disclosed, and has another arm against another specificity, or is present as a bivalent or multivalent molecule.

In addition to Antibody sequences the specific binding element may comprise other amino acids, the e.g. form a peptide or polypeptide, or in addition to the molecule ability To bind antigen to confer another functional property. For example, the specific binding element may be a marker, an enzyme or a fragment thereof include, etc.

The binding domain can be an entire VN domain or a part thereof, for which a germline segment or encodes a rearranged gene segment. The binding domain can a VL kappa domain or a VL lambda domain or parts thereof.

The binding domain may be a VH1, VH3 or VH4 germline sequence or a rearranged form include.

A specific binding element of the present invention may comprise a variable heavy chain (VN domain) region derived from the human germline DP47, whose sequence in FIG 1 (a) , Residues 1 to 98, is shown. The "DP" nomenclature is described by Tomlinson et al., (1992) The amino acid sequence of CDR3 may be Ser Leu Pro Lys The amino acid sequence of CDR3 may be Gly Val Gly Ala Phe Arg Pro Tyr Arg Lys His Glu Binding domain of a specific binding element according to the present invention comprise a VN domain comprising the amino acid sequences shown in 1 (a) for CGS1 and CGS2.

The binding domain may comprise a variable light chain ("VL" domain) region derived from the human germline DPL16, the sequence of which in FIG 1 (b) is shown as codons 1-90.

The VE domain may be a CDR3 sequence Asn Ser Ser Pro Val Val Leu Asn Gly Val Val include. The VL domain may be a CDR3 sequence Asn Ser Ser Pro Phe Glu His Asn Leu Val Val include.

Specific binding members of the invention may include functionally equivalent variants of the sequences described in U.S. Patent Nos. 5,236,866, 5,829,846, 5,829,877, 5,359,888, 5,348,877, and 5,734,686 1 in which, for example, one or more amino acids have been inserted, deleted, substituted or added, provided that a property set forth herein is retained. For example, the CDR3 sequence can be changed, or one or more changes can be made to the framework regions, or the framework can be replaced with another framework region or a modified form, provided the specific binding element binds ED-B.

One or more CDRs from a VL or VN domain of an antigen binding domain of an antibody disclosed herein may be in so-called "CDR transplantation" in which one or more CDR sequences of a first antibody are placed in a framework of sequences not derived from this antibody, eg from another antibody, as in EP-B-0239400 disclosed. CDR sequences for CGS1 and CGS2 are in 1 (a) and 1 (b) shown.

One specific binding element according to the invention may be one that has one herein for binding to fibronectin-ED-B described antibodies or scFv is competing. Competition between binding elements can be easily tested in vitro, e.g. by tying a specific one reporter molecule to a binding element in the presence of one or more others Unlabeled bonding elements can be detected to the To enable identification of specific binding elements the same epitope or an overlapping one Bind epitope.

One specific binding element according to the present Invention can be used in a method that causes the causation or enabling from binding of the specific binding element to its epitope. Binding may be the administration of the specific binding element to a mammal, e.g. a human or a rodent, e.g. a mouse, follow.

The The present invention provides the use of a specific binding element as above ready to use it as a diagnostic reagent for tumors to use. Evidence from experiments with animal models, below show that binding elements according to the present Invention for in in vivo tumor localization useful are.

preferred specific binding elements according to the present The invention includes those which, e.g. in a cryostat section, bind to human tumors that have an invasive and angiogenic Show phenotype and those which, e.g. in a cryostat section, on embryonic Tying fabric. Binding can be demonstrated by immunocytochemical staining become.

In a preferred embodiment does not or not significantly bind the specific binding element Tenascin, an extracellular Matrix protein.

In another preferred embodiment does not or not significantly bind the specific binding element normal human skin, e.g. in a cryostat section and / or demonstrated using immunocytochemical staining.

Further embodiments bind specific binding elements of the present invention not or not significant to one or more normal tissues (e.g., in a cryostat section and / or using immunocytochemical Coloring), selected liver, spleen, kidney, stomach, small intestine, Large intestine, ovary, uterus, bladder, pancreas, adrenals, skeletal muscle, Heart, lungs, thyroid and brain.

One specific binding element for ED-B can be used as an in vivo targeting agent to do so can be used to specifically detect the presence and location of tumors show that express fibronectin-ED-B or associated with it are. It can also be used as an imaging agent. The present The invention provides a method for detecting the presence of a Cell or tumor expressing fibronectin-ED-B or is associated with fibronectin-ED-B expression, which method comprises Contacting cells with a specific binding element and the determination of the binding of the specific binding element includes the cells. The method may be in vivo or in vitro a test sample of cells are removed from the body were.

The Reactivity of antibodies for one Cell sample can be determined by suitable means. Marking with individual reporter molecules is a possibility. The reporter molecules can directly or indirectly detectable and preferably measurable signals produce. The binding of reporter molecules can be direct or indirect, covalently, e.g. above a peptide bond, or non-covalently. Tie over one Peptide binding may be a consequence of recombinant expression of a Be a gene fusion for an antibody and a reporter molecule coded.

One preferred mode is via covalent binding of each antibody to an individual fluorescent dye, phosphorus or laser dye with spectrally isolated absorption or emission properties. Suitable fluorescent dyes include fluorescein, rhodamine, Phycoerythrin and Texas Red. Suitable chromogenic dyes include Diaminobenzidine.

Other Reporters include macromolecular colloidal particles or particulate matter, such as. Latex beads, the dyed, are magnetic or paramagnetic, and biological or chemical active agents that generate directly or indirectly detectable signals can, visually observed, electronically detected or otherwise recorded to become. These molecules can Enzymes that catalyze reactions that develop colors or change, or e.g. amendments cause electrical properties. They can be molecularly excitable, so that electron transitions between energy states lead to characteristic spectral absorptions or emissions. she can Chemical entities that are used in conjunction with biosensors be used. It can Detection systems for Biotin / avidin or biotin / streptavidin and alkaline phosphatase be used.

The Type of determination of binding is not a feature of the present Invention, and professionals are capable, according to their preferences and their general knowledge to select a suitable mode.

The signals generated from individual antibody-reporter conjugates can be used to derive quantifiable absolute or relative relative antibody binding data in cell samples (normal and test samples). In addition, a general nuclear stain, such as propidium iodide, can be used to specify the total cell population in a sample, allowing for the provision of quantitative ratios of individual cell populations compared to the total number of cells. Upon attachment of a radionucleotide, such as ¹²⁵ I, ¹¹¹ In or ^99m Tc, to an antibody, if this antibody preferentially locates in tumors rather than normal tissues, the presence of radiolabels in tumor tissues can be detected and quantified using a gamma camera , The quality of the tumor image obtained is directly correlated with the signal-to-noise ratio.

The antibody can used as diagnostic agents to newly vascularized Detect tumors, and can also e.g. used in modified form to be cytotoxic To supply funds or to trigger coagulation within new blood vessels, thereby causing the developing Tumor oxygen and nutrients be withdrawn, which is an indirect form of tumor therapy.

The The present invention also provides the use of a specific Binding element as described above ready to use as a therapeutic To use reagent, e.g. if it is coupled, bound or genetically engineered is produced as a fusion protein to an effector function too have. A specific binding element according to the present invention can Be used to a toxin, radioactivity, T cells, killer cells or other molecules to target a tumor expressing the antigen of interest or associated with it.

One Specific binding element can be used in treatment methods the administration of a specific binding element provided herein in pharmaceutical compositions containing a specific Binding element, and in uses of such a specific Binding element for the manufacture of a medicament for administration, for example in a process for the preparation of a medicament or a pharmaceutical composition containing the formulation of the specific binding element with a pharmaceutically acceptable carrier includes, used. Prepared compositions may be used Persons are administered. The administration is preferably in a "therapeutic effective amount, which is sufficient to provide benefits for one To give patients. Such an advantage can be at least alleviation be at least one symptom. The currently administered amount and the rate and timing of administration depend on the nature and severity its from what is being treated. The prescription of treatment, e.g. Decisions about the Dosage etc., is within the responsibility of general practitioners and other doctors. Suitable doses of antibodies are known in the art, see Ledermann et al. (1992), K.D. Bagshawe et al. (1991).

A Composition can be administered alone or in combination with other treatments, either simultaneously or in succession, dependent of the condition to be treated.

pharmaceutical Compositions according to the present Invention and for use according to the present Invention can additionally to the active ingredient a pharmaceutically acceptable excipient, carrier, buffer, Stabilizer or other materials known in the art. Such Materials should be non-toxic and the effectiveness of the active ingredient do not interfere. The exact nature of the carrier or other materials depends from the route of administration, this administration being oral or by Injection, e.g. intravenously.

Pharmaceutical compositions for oral administration may be administered in tablet, capsule, pulp Ver- or liquid form. A tablet may comprise a solid carrier, such as gelatin, or an adjuvant. Liquid pharmaceutical compositions include a liquid carrier such as water, petroleum, animal or vegetable oils, mineral oil or synthetic oil. It may be physiological saline, dextrose or other saccharide solution, glycols such as ethylene glycol, propylene glycol or polyethylene glycol.

to intravenous Infection or injection at the affected site may be the active ingredient in the form of a parenterally acceptable aqueous solution which is pyrogen-free is and has appropriate pH, isotonicity and stability. Professionals are capable of using e.g. isotonic carriers, like e.g. Sodium chloride injection, Ringer's solution and Ringer's lactate, appropriate solutions manufacture. Preservatives, stabilizers, buffers, antioxidants and / or other additives, if necessary, be included.

One specific binding element according to the present Invention can be made by expression from encoded nucleic acid become. nucleic acid the for encodes a specific binding element provided herein itself constitutes an aspect of the present invention, as well a method of producing the specific binding element, this process being expression from nucleic acid, the coded for includes. The expression can be cultivated appropriately of recombinant host cells comprising the nucleic acid, be achieved under suitable conditions.

The nucleic acid can for any of the amino acid sequences the antibody-antigen binding domains, the described herein, or any functionally equivalent Encode form. Changes can be made Nucleotide level by addition, substitution, deletion or insertion one or more nucleotides, these changes, dependent from the degeneration of the genetic code, also at the amino acid level can reflect or not.

systems for cloning and expression of a polypeptide in a variety of different host cells are known. Suitable host cells include bacteria, mammalian cells, Yeast and baculovirus systems. Mammalian cell lines available in the field for expression of a heterologous polypeptide include Chinese hamster ovary cells, HeLA cells, Baby hamster kidney cells and many others. A common preferred bacterial host is E. coli.

The Expression of antibodies and antibody fragments in prokaryotic cells, e.g. E. coli is good in the art established. For For a report see, for example, Plückthun (1991). Expression in Eukaryotic cells in culture is considered an option for production of a specific binding element, for recent reports see e.g. reef (1993); Trill et al. (1995).

It can chosen suitable vectors or produced, containing suitable regulatory sequences, including Promoter sequences, termination sequences, polyadenylation sequences, enhancer sequences, Marker genes and other suitable sequences. Vectors can, je as required, plasmids, such as viral, e.g. phages, or phagemids be. For for further details see e.g. Molecular Cloning: A Laboratory Manual: 2nd edition, Sambrook et al. (1989), Cold Spring Harbor Laboratory Press. Many known methods and regulations for manipulation of nucleic acid, e.g. for the preparation of nucleic acid constructs, mutagenesis, Sequencing, introduction of DNA in cells and gene expression and analysis of proteins in Short Protocols in Molecular Biology, 2nd Ed., Ausubel et al. Ed., John Wiley & Sons (1992) described in detail. The disclosures of Sambrook et al. and Ausubel et al. are incorporated herein by reference.

Therefore another aspect of the present invention is a host cell ready, the nucleic acid as disclosed herein. Another aspect provides a methodology for introducing a such nucleic acid in a host cell. The introduction can be by any Procedure done. For eukaryotic cells can suitable methods calcium phosphate transfection, DEAE-dextran, electroporation, liposome-mediated transfection and transduction using Retroviruses or other viruses, e.g. Vakzinia, or for insect cells Baculovirus include. For bacterial cells can suitable procedures calcium chloride transformation, electroporation and transfection using a bacteriophage.

On the introduction can the induction or enabling of expression from nucleic acid follow, e.g. by culturing host cells under conditions for the expression of the gene.

In one embodiment, the nucleic acid of the invention is integrated into the genome (eg, chromosome) of the host cell. Integration can be promoted by incorporation of sequences prepared according to standard Ver drive to promote recombination with the genome.

To the production of a specific binding element may be, for example in any of the ways disclosed herein, such as e.g. in the formulation a pharmaceutical or diagnostic product, such as. a set as described in addition to the specific binding element one or more reagents for determining the binding of the element to cells.

Further Aspects and embodiments The invention is known to those skilled in the art. To fully understand the present invention, the following examples are provided, for illustrative purposes only and not as a limitation should.

1 shows aligned amino acid sequences of the VH and VL of the scFvs CGS-1 and CGS-2. 1 (a) shows VN sequences, 1 (b) shows VL sequences. CDRs (1, 2 and 3) are indicated. The most homologous human germline VH to both scFvs is the DP47 segregation of the VH3 family; the VL segment of both clones is DPL16, the light chain used to form the original scFv library (Nissim et al., 1994). Residues that distinguish the two clones are underlined.

2 : 2A shows a model of the domain structure of a human FN subunit. The IIICS, ED-A, ED-B regions are indicated with variability due to alternative splicing of the FN pre-mRNA. The figure also shows the internal homologies as well as the major thermolysin digestion products containing ED-B (Zardi et al., 1987). 2 B indicates 4-18% SDS-PAGE of plasma and WI38VA-FN and their thermolysin cleavage products stained with Coomassie blue, as well as immunoblots using BC-1, IST-6, CGS-1 and CGS-2 were probed. Undigested (lane 1) and plasma FN digested using thermolysin with 1 μg / mg FN (lane 3) and 10 μg / mg FN (lane 4). Undigested (lane 2) and WI38VA-FN digested with thermolysin at 1 μg / mg (lane 5), 5 μg / mg (lane 6) and 10 μg / mg (lane 7) FN. The numbers on the right indicate the major thermolysin digestion products found in 2A are shown. The values on the left indicate the molecular weight standards in kilodaltons (kD).

3 : 3A Figure 4 shows the FN type III repeat sequences contained in the fusion and recombinant proteins expressed in E. coli and the reactivity of these proteins with CGS-1 and CGS2 and with the MAbs BC-1 and IST. 6th 3B shows a Coomassie blue stained gel and then the immunoblots probed with CGS-1, CGS-2, BC-1, IST-6. The numbering of the lanes corresponds to that of the peptide constructs in the upper part of the figure. The values on the left indicate the molecular weight standards in kD.

4 : Infrared Mouse Imaging Device; the mouse imager used for these targeting experiments consisted of a black, nonfluorescent box equipped with a tungsten halogen lamp, excitation and emission filters specific to the CY7 infrared fluorophore, and a computer controlled 8-bit monochrome CCD Camera was equipped.

5 : Targeting Fluorescently Labeled Antibody Fragments to Murine F9 Teratocarcinoma Using the Monomeric scFv (CGS-1) and Dimeric scFv (CGS-1) ₂ Targeted to B-FN. Dimeric scFv (D1.3) ₂ with binding specificity for lysozyme was used as a negative control.

6 .: Targeting fluorescent-labeled antibody fragments to the murine F9 teratocarcinoma using affinity-matured scFv (CGS-2) and lower affinity scFv (28SI) aligned with the same epitope of B-FN. Targeting is detected for both large tumors (about 0.6 g) covered with a black crust for 48 h, which partially darkens the imaging, and small tumors (about 0.2 g).

List of examples

• Example 1 - Isolation of human scFvs used for the ED-B domain are specific to human FN.
• Example 2 - Affinity maturation of human scFvs used for the ED-B domain are specific to human FN.
• Example 3 - Specificity of affinity matured scFvs for ED-B-containing fibronectins.
• Example 4 - Use of affinity matured Anti-ED-B scFvs for immunocytochemical staining of human and mouse tumor sections.
• Example 5 - Use of affinity matured anti-ED-B scFvs for in vivo human tumor targeting ren.
• Example 6 - Aiming xenografted murine F9 teratocarcinoma in nude mice.

Example 1 - Isolation of human scFvs for the ED-B domain are specific to human FN

A human scFv phage library (Nissim et al. (1994)) has become the Selection of recombinant antibodies used. Two different Forms of the ED-B isoform were used as antibody source used for selection, and in both cases the isoform was recombinant human protein.

recombinant FN peptides containing Type III repeats 2-11 (B-) and 2-11 (B +) were expressed in Escherichia coli.

One Construct was constructed using FN cDNA from clones pFH154 (Kornbluth et al., 1985), λF10 and λF2 (Carnemolla et al. (1989)). The cDNA constructs that are concerned with the Bases 2229-4787 (Kornblihtt et al. (1985)) Use of the QlAexpress set from Qiagen (Chatsworth, California) inserted into the vector pQE-3/5. Recombinants FN-III 2-11 (B-) and (B +) were used by immunoaffinity chromatography of MAK 3E3 (Pierschbacher et al. (1981)) purified on Sepharose-4B conjugated (Pharmacia). DNA fragments for the production of the recombinant FN fragments carrying the type III homology repeats 7B89, 789, Containing ED-B and FN-6 were detected by polymerase chain reaction (PCR) amplification using UltMa DNA polymerase (Perkin Elmer), using cDNA from clones FN-2-11 (B +) and FN 2-11 (B +) produced as matrices. There were primers designed to the cloning of PCR products into pQE-12 using the QIAexpress kit (Qiagen). They were subsequently transformed into E. coli and expressed. All cDNA clones were made using a Sequenase 2.0 DNA sequencing kit (USB) sequenced.

recombinant Proteins were as indicated from the manufacturer (Qiagen) by means of Ni-NTA chromatography (IMAC) Use of a hexahistidine tag at the carboxy terminus of the FN fragments cleaned. The ED-B-βGal fusion protein was prepared by cloning ED-B cDNA into the λgt11 bacteriophage vector, around the clone λED-B to surrender. Clone λchFN60 (which contained part of the ED-B sequence) was used as a fusion protein from the cloned chick FN pchFN60 (Norton et al., 1987).

to Selection of the human scFv phage library, were for each of the two different recombinant antigens (7B89 and ED-B) performed three panning rounds. The antigens were both over Night at 50 μg / ml in PBS (20 mM phosphate buffer, 0.15 M NaCl, pH 7.2) layered on immunotubes (Nunc; Maxisorp, Roskilde, Denmark). The first antigen was the recombinant FN fragment 7B89 in which the ED-B domain flanked by adjacent type III FN homology repeats, this became overnight at 4 ° C coated. The second antigen used was recombinant ED-B (Zardi et al., 1987) with a carboxy-terminal hexahistidine tag; this protein did not contain any lysine residues, so the terminal Amino group of the first amino acid for site-specific covalent immobilization of ED-B on reactive ELISA plates (Nunc, Covalink) available is. The coating was over Night at room temperature.

To Three rounds of panning were eluted phages in HB2151 E. coli cells infected and plated as described (Nissim et al. (1994)). After each round of selection, 95 ampicillin-resistant individuals were obtained Screened clones to identify antigen-specific scFvs by ELISA. Clones, which are the strongest ELISA signals on the antigens that were used for panning were selected for further analysis and for affinity purification. It It has also been shown that these clones undergo immunocytochemical staining specific staining from sections of glioblastoma multiforme and breast tumors revealed as described in more detail in Example 4.

Example 2 - Affinity maturation of human scFvs for the ED-B domain are specific to human FN

Clones 35GE (from selection with 7B89) and 28SI (from selection with the ED-B domain alone) were selected as candidate antibodies for affinity maturation. To diversify the light chain as a means of improving affinity, the inventors discovered a simple affinity maturation strategy based on randomizing the central six residues (DSSGNH) of the light chain CDR3 using degenerate oligonucleotides and PCR ( 1 ), providing a potential sequence diversity of 20 ⁶ = 6.4 × 10 ¹ . This region (along with the heavy chain CDR3) is at the center of the antigen binding site (Padlan (1994)). The inventors also mutated the arginine residue just before the six residues to serine to avoid the possibility of electrostatic effects that might dominate the selection.

Plasmids from a single bacterial colony expressing the "strain" ScFV fragment were amplified by PCR with the primers LMB3 (5 'CAG GAA ACA GCT ATG AC 3') and CDR3-6 VL-FOR (5 'CTT GGT CCC TCC GCC GAA TAC CAC MNN MNN MNN MNN MNN AGA GGA GTT ACA GTA ATA GTC AGC CTC 3 ') (94C [1'] - 55C [1 '] - 72C [1'30 "], 25 cycles, see Marks et al. (1991) for Buffers and Conditions). The resulting product was gel purified (to remove traces of the plasmid containing the original scFv gene) and used as a template for a second amplification step with primers LMB3 and J1-Not-FOR (5 'ATT GCT TTT CCT TTT TGC GGC CCC GCC TAG GAC GGT CAG CTT GGT CCC TCC GCC 3 ') (94C [1'] - 55C [1 '] - 72C [1'30 "], 25 cycles) The crude PCR product, which was used as a single band run on agarose gel with correct molecular weight, was purified from the PCR mix directly using spin-bind (FMC, Rockland, Maine, USA), digested twice with Nco1 / Not1, and into gel-purified, Nco1 / Not1 digested phagemid pHNE1 (Hoogenboom et al., 1991)), which contained a cloned Nco1 / Not1 insert to allow the separation of a double digested from a single digested vector The vector was digested with a Qiagen plasmid maxi-prep kit (Chatsworth, California, USA) About 5 μg of digested plasmid and insert were ligated in the ligation mixture which was extracted once with phenol, once with phenol / chloroform / isoamyl alcohol (25: 25: 1), then ethanol precipitated using glycogen (Boehringer, Mannheim, Germany) and dried at Speed-vac. The pellet was resuspended in 20 μl of water and electroporated into electrocompetent TG1 E. coli cells (Gibson (1984)). The inventors typically used electrocompetent cells with a titer of 10 ⁹ transformants / μg when glycerol strains were used or 10 ¹⁰ transformants / μg in freshly prepared electrocompetent cells. This typically gave> 10 ⁷ clones by the method reported herein.

The maturation library was then processed as the Nissim library (Nissim et al., 1994) to produce phage particles eligible for round-robin selection using 7B89 (10 μg / ml) as an antigen, followed by a round of kinetic selection (Hawkins et al., 1992)). This selection step was performed by incubating biotinylated 76B89 (10 nm) with the phage suspension (about 10 ¹² transfer units) in 2% milk PBS (2% MPBS) from the first round of selection for 56 minutes, then non-biotinylated 7B89 (1 μM ) and the competition continued for 30 minutes. 100 μl of streptavidin-coated Dynabeads (Dynal: M480) preblocked in 2% MPBS were then added to the reaction mixture, mixed for 2 minutes and then captured on a magnet and washed 10 times with alternating washings (PBS + 0.1% Tween). 20) and PBS. The phage was eluted from the beads with 0.5 ml of 100 mM triethylamine. This solution was then neutralized with 0.25 ml of 1 M Tris, pH 7.4, and used to infect exponentially growing HB2151 cells (Nissim et al., 1994). Ninety-five individual ampicillin-resistant colonies were used to produce scFv-containing supernatants (Nissim et al., 1994) which were screened by ELISA, immunohistochemistry, and BlAcore to identify the best binders. They were then subcloned between Sfi1 / Not1 sites of the pDN268 expression vector (Neri et al., (1996)), which attaches a phosphorylatable label, the FLAG epitope, and a hexahistidine label to the C-terminal arm of the scFv.

Individual colonies of the relevant antibodies subcloned into pDN268 were cultured at 37 ° C in 2xTY containing 100 mg / l ampicillin and 0.1% glucose. When the cell culture reached an OD ⁶⁰⁰ of 0.8, IPTG was added to a final concentration of 1 mM and growth was continued for 16-20 hours at 30 ° C. After centrifugation (GS-2 Sorvall rotor, 7000 rpm, 30 minutes), the supernatant was filtered, concentrated and loaded using a tangential minisette (Filtron) flow-through device against charge buffer (50 mM phosphate, pH 7.4, 500 mM NaCl, 20 mM imidazole). The resulting solution was loaded onto 1 ml Ni-NTA resin (Qiagen), washed with 50 ml charge buffer and eluted with elution buffer (50 mM phosphate, pH 7.4, 500 mM NaCl, 100 mM imidazole). The purified antibody was analyzed by SDS-PAGE (Laemmli (1970)) and dialyzed against PBS at 4 ° C. Purified scFv formulations were further processed by gel filtration using an FPLC device equipped with an S-75 column, as it is known that multivalent scFv fragments are affected by avidity effects (Nossom et al. (1994), Crothers and Metzger (1972)) show an artificial good binding to BlAcore (Jonsson et al. (1991)). The antibody concentration of FPLC-purified monomer fractions was determined spectrophotometrically, assuming an absorbance at 280 nm of 1.4 units for a 1 mg / ml scFv solution.

The binding of monovalent scFvs at various concentrations in the 0.1-1 μM range in PBS was using the following antigens on a BIAcor (Pharmacia Biosensor) instrument: (i) 1000 resonance units (RU) of biotinylated recombinant FN fragment 7B89 immobilized on a streptividin coated chip, then specifically bound by 250 RU scFv has been; (ii) 200 RU of recombinant ED-B chemically immobilized at the N-terminal amino group, which was then specifically bound by 600 RU scFv; (iii) 3500 RU of ED-B-rich fibronectin WI38VA (see Example 3), which was then specifically bound by 150 RU scFv. A kinetic analysis of the data was performed according to the manufacturer's instructions. Based on the qualitative BlAcore analysis of antibody-containing supernatants, an affinity matured version of each scFv clone was selected: clone CGS-1 by selection with the 78B9 fragment and cgS-2 by selection with recombinant ED-B-FN fragment. The association rate constants (k _on ) and dissociation rate constants (k _off ) are reported in Table 1 together with the calculated equilibrium dissociation constants (Kd) of both the scFvs and the original 28SI clone. Although both CGS-1 and CGS-2 clones possess nanomolar Kds, clone CGS-2 showed the best improvement over its parent clone, resulting in a single assay for all three proteins tested on the sensor chip (Table 1) Kd of 1 nM (improved over 110 nM). The improvement was mainly due to a slower kinetic dissociation constant (10 ^-4 s ^-1 ) as measured with monomeric antibody formulations (not shown).

The Maturation strategy seems to be general and yielded affinity-enhanced antibody against maltose binding protein, cytochrome C, the extracellular domain of murine Endoglin (D.N., L. Wyder, R. Klemenz), cytomegalovirus (A.P.G. Neri, P.N. R. Botti), the nuclear tumor marker HMGI CP protein (A. P., P. Soldani, V. Giancott, P.N.) and the placental alkaline Phosphatase as an ovarian tumor marker (M. Deonarain and A. A. Epenetos). The strategy therefore seems to be at least as effective as Other Maturation Strategies (Marks et al., 1992, Low et al., 1996) and delivers antibodies with similar ones affinities like those of very big ones Phage antibody libraries resulting (Griffiths et al., 1994, Vaughan et al., (1996)).

The affinity matured clones CGS-1 and CGS-2 were sequenced and aligned with a database of human germ-line antibody V genes (V-BASE), then translated using MacVector software. The VR gene of both clones was the most homologous to the human germ line DP47 (VH3), and in addition each clone had a different VH CDR3 sequence ( 1 ). The VR gene of both clones was the DPL16 germline used to make the human synthetic scFv repertoire described by Nissim et al. (1994). VL-CDR3 sequences differed on four out of six randomized residues ( 1b ).

Example 3 - Specificity of affinity matured scFvs for ED-B-containing fibronectins

The immunoreactivity of the two affinity-matured scFvs CGS-1 and CGS-2 was initially assessed using ELISA and assessed directly with MAK BC-1 (which recognizes the B-FN isoform) and MAK IST-6, which recognizes only FN isoforms lacking ED-B (Carnemolla et al. (1989) (1992)). , compared. The characterization of these MAbs has previously been reported (Carnemolla et al., (1989) (1992)). The fine specificity analysis was then carried out using a large group of FN fragments obtained by thermolysin treatment and recombinant fusion proteins.

The antigens used for ELISA and immunoblotting were prepared as follows. FN was purified from human plasma and from the conditioned medium of the WI38VA13 cell line as previously reported (Zardi et al., 1987). Purified FNs were probed with thermolysin (Protease Type-X, Sigma Chemical Co.) as described by Carnemolla et al. (1989) reported digested. Native FN fragments with 110 kd (B) and 120 kd (B +) (see 2 ) were purified from FN digestion as previously reported (Borsi et al. (1991)). The large isoform of tenascin-C was prepared as described by Saginati et al. (1992) purified, recombinant proteins were expressed and purified as described in Example 1. SDS-PAGE and Western blotting were performed as described by Carnemolla et al. (1989).

All antigens used in the ELISA were diluted in PBS to between 50-100 μg / ml and coated at 4 ° C overnight on immunoplate wells (Nunc, Roskilde, Denmark). Unbound antigen was removed with PBS and the plates were then blocked for 2 h at 37 ° C with PBS containing 3% (w / v) bovine serum albumin (BSA). This was followed by four washes with PBS containing 0.05% Tween 20 (PEST). Antibodies were then allowed to bind at 37 ° C for 1.5 h, scFvs were preincubated with antiserum directed against the taq sequence: MAK M2 [Kodak, New Haven, Connecticut] for FLAG labeling or 9E10 [ATCC , Rockville, Maryland] for the myc mark. The control antibodies tested were MAbs BC-1 and IST-6. After four washes with PEST, the plates were diluted for 1 hour at 37 ° C with 1: 2000 diluted (in PEST + 3% BSA) biotinylated goat anti-mouse IgG (Bio-SPA Division, Milan, Italy). The washes were repeated and a complex of streptavidin and biotinylated alkaline phosphatase (Bio-SPA Division, Milan, Italy) was added for 1 h at 37 ° C (diluted 1: 800 in PEST containing 2 mM MgCl ₂ ). The reaction was developed using phosphatase substrate tablets (Sigma) in 10% diethanolamine, pH 9.8, and the optical density was read at 405 nm. The results are shown in Table 2 below. TABLE 2 CGS-1 CGS-2 BC-1 IS-6 Plasma FN 0.07 0.04 0.09 1.73 WI38VA TN 1.16 0.72 1.20 1.12 n110 kD (B) 0.03 0.01 0.05 1.20 n120 kD (B +) 0.82 0.81 1.20 0.02 rec FN7B89 1.11 1.02 1.02 0.01 rec FN789 0.01 0.01 0.05 1.25 rec ED-B 1.21 1.32 0.15 0.04 rec FN-6 0.01 0.01 0.08 0.03 tenascin 0.01 0.02 0.06 0.02

Immunoreactivity of scFv and monoclonal antibodies with fibronectin-derived antigens measured by ELISA. The values represent the OD measured at 405 nm after subtraction of the background signal has been. The data are the means of four experiments, the one maximum standard deviation of 10%.

The identity of the different forms of fibronectin used in this experiment was as follows: plasma FN = human plasma fibronectin; WI38-VA-FN = fibronectin from supernatants of SV40-transformed fibroblasts (Zardi et al., 1987); n110kd = thermolysin-treated FN domain 4 without ED-B; n120kd = with Thermolysin treated FN domain 4 containing ED-B; rec FN7B89 = ED-B domain flanked by adjacent ones Type III FN homology repeats; rec FN789 = Type III FN homology repeats with an ED-B domain; rec ED-B = recombinant ED-B alone; rec FN6 = recombinant FN domain 6.

Both CGS-1 and CGS-2 recognized the recombinant ED-B peptide, as well as any native or recombinant FN fragments containing the ED-B sequence, while they did not bind to any FN fragments to which ED- B was missing. Furthermore, CGS-1 and CGS-2 did not react with tenascin (comprising fifteen type III homology repeats: Siri et al (1991)) and plasma Fn incorporated in thermolysin digestion contained no detectable levels of the ED-B sequence (Zardi et al., (1987)). In contrast, CGS-1 and CGS-2 reacted strongly with FN purified from the SV40-transformed WI378VA cell line. About 70-90% of FN molecules from this cell line contain ED-B as demonstrated by thermolysin digestion and S1 nuclease experiments using purified FN and total RNA made from the cell line (Zardi et al (1987); Borsi et al. (1992)). The specificity of scFVs for the ED-B component of FN was further demonstrated by the use of soluble recombinant ED-B to inhibit binding of CGS-1 and / or CGS-2 to FN on WI38VA cells (data not specified).

The Confirm data, that CGS-1 and CGS-2 only react specifically with FN derivatives, which is the ED-B domain contain. They both show the same reactivity as MAK BC-1, with the exception the case of recombinant ED-B not recognized by BC-1. The intensity the ELISA signals obtained compared to the MAK controls shows the high specificity of the two scFvs for ED-B-containing antigens.

The specificity of CGS-1 and CGS-2 was further examined on immunoblots using FN from plasma and WI38VA cells and thermolysin digests thereof. After thermolysin digestion, FN generated from WI38VA cells (majority of which contained ED-B) a 120 kd fragment (containing ED-B) and a smaller 110 kd fragment lacking ED-B ( 2A ; Zardi et al. (1987)). Further digestion of the 120 kd domain produces two fragments; an 85 kd fragment containing almost all of the ED-B sequence at its carboxy-terminus and a 35 kd sequence ( 2A , Zardi et al., (1987)).

On the left side of 2 B is a Coomassie blue stained gel of protein fractions analyzed by immunoblotting. Plasma FN (lane 1) and thermolysin digests of the protein (lane 3 containing the 110 kd protein and lane 4 containing digested 110 kd protein) were not recognized by CGS-1 and CGS-2. In contrast, ED-B-rich FN was recognized from WI38VA cells both intact (lane 2) and after increasing thermolysin digestion (lanes 5, 6, and 7) from both scFv fragments. The smallest fragment derived from FN that could be specifically recognized by CGS-1 was the 120 kD protein (which spanned type III repeats 2-11 (bounds included)), while CGS-2 was the 85 kD Fragment that spanned repeats 2-7, in addition to the N-terminus of ED-B ( 2 B ; Zardi et al. (1987)). These results indicate that the two scFvs respond to different epitopes within the ED-B sequence. The binding of CGS-2 to the 85 kd domain shows that the epitope for this clone lies in the amino terminus of ED-B. In contrast, when the 120 kD domain is digested to 85 kD, the loss of CGS-1 binding reveals that it recognizes an epitope that is further toward the carboxy terminus of the ED-B molecule.

The fine specificity of CGS-1 and CGS-2 was further investigated by immunoblotting using recombinant FN fragments and fusion proteins with or without the ED-B sequence. The FN fusion proteins were prepared as described by Carnemolla et al. (1989). The results of these experiments are in 3 for the association of the schematic diagram with the structure of the domains of human FN see Carnemolla et al. (1992). The binding profiles obtained essentially confirmed what had previously been carried out by ELISA and immunoblots on purified FN and proteolytic cleavage products: CGS-1 and CGS-2 reacted strongly with ED-B-containing FN fragments (lanes 2 and 4), but did not show any Reactivity with FN sequences lacking ED-B (lanes 1 and 3). CGS-1 did not react with either human (lane 5) or chicken (lane 6) ED-B fusion protein, whereas CGS-2 reacted strongly with both fragments ( 3 ). This result may reflect certain conformational limitations of the epitope in ED-B-containing FN recognized by CGS-1; for example, it is possible that the epitope is sensitive to denaturation or it may not be correctly displayed when analyzed by SDS-PAGE is fractionated and transferred to a solid support such as nitrocellulose.

Taken together, these results show that CGS-1 and CGS-2 are located in regions which differ from each other and other than the ED-B structure strong and specific detected by the MAK BC-1 Bind ED-B-containing FNs.

Example 4 - Use of affinity matured Anti-ED-B scFvs for immunocytochemical staining of human and human Mouse tumors

CGS-1 and CGS-1 have both been used to immunolocalize ED-B containing FN molecules in various normal and neoplastic human tissues. Skin was selected as the normal tissue since it is known that the B-FN isoform is expressed in macrophages and fibroblasts during cutaneous wound healing (Carnemolla et al., 1989, Brown et al., 1993). The two selected human tumors have previously been analyzed for the specificity of staining with anti-fibronectin MAbs: Glio Blastoma multiforme has been studied in detail because endothelial cells in the vessels of this tumor are in a highly proliferative state with enhanced angiogenic processes involving the expression of B-FN isoforms (Castellani et al., 1994). Further, studies using a diverse group of normal, hyperplastic and neoplastic human breast tissue have provided further evidence for correlations between angiogenesis and B-FN expression (Kaczmarek et al., 1994).

For the herein described experiments is the immunohistochemical staining of CGS-1 and CGS-2 with that of MAK BC-1 (which recognizes the B-FN isoform) and other MAKs known to be they either on all known FN isoform variants (IST-4) or respond only to FN isoforms lacking ED-B (IST-6). It is previously reported on the characterization of all these control antibodies (Carnemolla et al. (1989) (1992)).

normal and neoplastic tissues were obtained from samples used in surgical Interventions were taken. It has already been stated that the preparation and fixation of tissues for the exact and sensitive detection of FN-containing molecules is (Castellani et al. (1994)). For immunohistochemistry, 5 μm thick cryostat sections were used air-dried and fixed in cold acetone for ten minutes. Immunostaining was using a streptavidin-biotin-alkaline phosphatase complex stain kit (Bio-SPA Division, Milan, Italy) and naphthyl AS-MX phosphate and Fast Red TR (Sigma) performed. Gills hematoxylin was as a counterstain followed by inclusion in glycerol (Dako, Carpenteria, California) as previously described by Castellani et al. (1994). Around the specificity Furthermore, in experiments to analyze in which a positive staining of Tissues was obtained, the specificity for ED-B by preincubation of antibodies with the recombinant ED-B domain, obtained following the detection described above.

The Results of these experiments generally showed that both CGS-1 as well as CGS-2 react with the same histological structures like the MAK BC-1. That with skin using CGS-1, CGS-2 and BC-1 color pattern reflects the absence of ED-B in FN, the in the dermis was expressed. In the staining of invasive ductal carcinoma sections CGS-1, CGS-2 and BC-1 showed a limited distribution of staining, the was confined to the border between the neoplastic cells and the stream. This is consistent with the fact that, though the entire FN is homogeneously distributed throughout the tumor stream, expression B-FN is limited to certain regions, and it is these Regions previously successful (in 95% of cases) in invasive ductal Carcinoma using the MAK BC-1 have been isolated (Kaczmarek et al. (1994)).

It are previous findings in the staining of BC-1 of the tumor glioblastoma multiforme has been proven. Castellani et al. (1994) a typical coloring pattern of glomerular-like vascular structures observed, and in the experiments of the inventors it has been shown that CGS-1 and CGS-2 give qualitatively identical results.

It However, there is one important difference between CGS-1 and CGS-2 and the MAK BC-1: It has been shown that the two human scFvs on chicken and mouse B-FN bind while BC-1 is strictly human-specific. CGS-2 reacted with chicken embryos (Data not shown), and both CGS-1 and CGS-2 responded with mouse tumors.

It is also CGS-1 staining of vascular Structures shown on sections of murine F9 teratocarcinoma Service. On the other hand, all tested normal mouse tissues (liver, Spleen, kidney, stomach, small intestine, Large intestine, ovary, uterus, bladder, pancreas, adrenal glands, skeletal muscle, Heart, lungs, thyroid and brain) a negative staining reaction with CGS-1 and CGS-2 (data not shown). The in the F9 teratocarcinoma sections colored Structures proved to be ED-B specific by the recombined ED-B domain was used to complete the obtained color inhibit (data not specified)

Example 5 - Use of affinity matured Anti-ED-B scFvs for in vivo targeting of human tumors

The human melanoma cell line SKMEL-28 was used to develop xenografted tumors in 6-10 week old nude mice (Balb-C or MF-1, Harlan, UK) by injecting 1 x 10 ⁷ cells / mouse into a flank subcutaneously were.

Mice bearing tumors were injected into the tail vein with 100 μl of 1 mg / ml scFv ₁ -Cy 7 ₁ solution PBS injected when the tumors had reached a diameter of about 1 cm.

Labeling of recombinant antibodies with CY7 was made to be 1 by addition of 100 μl of 1 M sodium bicarbonate, pH 9.3, and 200 μl of CY7-bis-OSu (Amersham, cat. # PA17000; 2 mg / ml in DMSO) ml antibody solution in PBS (1 mg / ml). After 30 minutes at room temperature, 100 μl of 1 M Tris, pH = 7.4 were added to the mixture and the labeled antibody was purified using single use PD10 columns (Pharmacia Biotech, Piscataway, New Jersey, USA) with PBS equilibrated, separated from unreacted dye. The eluted green antibody fractions were concentrated to about 1 mg / ml using Centricon 10 tubes (Amicon, Beverly, Massachusetts, USA). The labeling ratio achieved was generally close to 1 molecule of CY7 per 1 molecule of antibody. This was assessed spectroscopically with 1 cm cuvettes, assuming that 1 mg / ml antibody solution gives an absorbance of 1.4 units at 280 nm, that the molar extinction coefficient of CY7 at 747 nm is 200,000 (M ^-1 cm ^-1 ) ^{. 1} ), neglecting CY7 absorption at 280 nm. Immunoreactivity of the antibody samples after labeling was detected either by band shifting (Neri et al. (1996b)) by affinity chromatography on an antigen column or by BIA core analysis. The mice were imaged under anesthesia by inhaling an oxygen / fluorothane mixture with a home-made mouse imaging device at regular intervals. For each sample, two to eight animals were studied to ensure reproducibility of the results. The procedures were performed according to the license of the UK project "Tumor Targeting" awarded to D. Neri (UK PPL 80/1056).

The Infrared mouse imaging device was as a modification of the photodetection system of Folli et al. (1994) which allows the use of the infrared fluorophore CY7. Infrared lighting was selected to achieve better tissue penetration. The fluorescence of CY7 (> 760 nm) is for humans invisible and requires the use of a computer-controlled CCD camera. The mouse imaging device existed from a black colored, opaque Box equipped with a 100 W tungsten halogen lamp, those with a 50 mm diameter excitation filter equipped was that specific to CY7 (Chroms Corporation, Brattleboro, Vermont, USA; 673-748 nm). The resulting illumination beam is in good approximation over a area of 5 × 10 cm homogeneous, in which the mouse was placed for imaging. fluorescence was detected by an 8-bit monochrome Pulnix CCD camera, the equipped with a C-mount lens and a 50 mm emission filter was (Chroms Corporation, Brattleboro, Vermont, USA, 765-855 nm) and over the ImageDOK system was connected (Kinetic Imaging Ltd., Liverpool, UK). This system consists of a computer that comes with a frame grabber and software for receiving and integrating successive ones Equipped pictures is. Three consecutive images obtained in 50 ms were each typically used for averaging, wherein this number for the series of images of an animal was kept constant to one direct comparison of tumor targeting at different times to enable. Images in TIFF format were then using the program Graphics Converter converted to PICT files and using the program MacDraw Pro designed with a Power Macintosh 7100/66 computer.

A schematic sketch of the construction of this device is in 4 shown.

The Experiments showed that both scFvs localized in the tumor, when visualized at the macroscopic level.

The microscopic demonstration of targeting on neovasculature of developing tumors were made with the two anti-EDB scFvs in detail as follows.

Nude mice and / or SCID mice either a human xenografted SKMEL-28 melanoma or a mouse F9 teratocarcinoma in a flank either unlabelled scFV fragments with the FLAG tag or injected biotinylated scFv fragments.

The Mice were killed at different times after the injection, it tumor and non-tumor sections were obtained, which were then Immunohistochemistry protocols stained using either the anti-FLAG-M2 antibody (Kodak 181) or streptavidin-based Detection reagents were used. Optimal targeting will generally achieved 12 hours after injection. It was shown, that both CGS1 and CGS2 transplanted the neovasculature of both the xenograft human tumor as well as murine teratocarcinoma.

Example 6 - Targeting xenografted murine F9 teratocarcinoma in nude mice

The inventors developed solid tumors in the flank of nude mice by subcutaneous injection of 4 x 10 ⁶ murine F9 teratocarcinoma cells. This tumor grew very rapidly in mice, reaching 1 cm in diameter about one week after injection, and is highly vascularized. To illustrate the targeting to the antibodies, the inventors used a modification of the photodetection method of Folli et al. (1994), which showed a kinetic evaluation of tumor targeting and antibody clearance on the same animal, as illustrated at various times, as described in detail above (see 4 ).

To the Targeting the tumor and facilitating the detection of antibodies were scFv (CGS-1), scFv (CGS-2) and anti-lysozyme scFv (D1.3) (McCafferty et al. (1990)) provided with a homodimerization label (Pack et al. (1993)), by the antibodies subcloned into the Sfi1 / Not1 sites of the expression vector pGIN50 were. This vector is a derivative of pDN268 (Neri et al. (1996b)), in which the His6 sequence of the label is represented by the following sequence is replaced: GGC LTD TLQ AFT DQL EDE KSA LQT EIA HLL KEK EKL EFI LAA H, which contains a cysteine residue and the amphipathic helix of the Fos protein for covalent homodimerization of antibody fragments (Abate et al. (1990)). It did not become a complete covalent dimerization achieved: about 30-50% of the antibody fragments consisted of covalently bonded dimers.

Antibody fragments were purified by affinity chromatography on columns prepared by binding chicken egg lysozyme (D1,3) or 7B89 (anti-ED-B antibody; Carnemolla et al. (1996)) to CNBr-activated Sepharose (Pharmacia Biotech, Piscataway, New Jersey, USA). Supernatants were loaded onto the affinity supports, which were then washed with PBS and with PBS + 0.5 M NaCl and eluted with 100 mM Et ₃ N. The antibodies were then dialyzed against PBS.

The antibodies were labeled as described above and then injected into the tail vein of tumor bearing mice with 100 μl of 1 mg / ml scFv ₁ -Cy7 ₁ solution in PBS when the tumors reached a diameter of about 1 cm.

As in 5 As shown, scFv (CGS-1) localized to the tumor for up to three days, although there was rapid clearance from the tumor during this period. However, there was also some staining of the femur. The targeting performance of CGS-1 to the tumor was dramatically improved by the introduction of an amphipathic helix containing a cysteine residue at the C-terminus to promote antibody dimerization (Pack et al., 1993). Indeed, localization of dimeric scFv (CGS-2) ₂ did not appear to decrease significantly for 24 to 72 hours. In contrast, a negative control (the dimeric antibody scFv (D1,3) ₂ , anti-lysozyme antibody) showed rapid clearance and no detectable localization on the tumor or femur.

ScFv (28SI) showed poor tumor targeting after 6 hours (not shown), but none could be detected after 24 hours or later ( 6 ). Affinity maturation resulted in greatly improved targeting, whereby scFv (CGS-2) effectively targeted small and large F9 tumors, whether as a monomer ( 6 ) or dimer (not shown). After two days, the percent injected dose of antibody per gram of tumor was found to be about 2 for the scFv (CGS-2) monomer and 3-4 for the scFv (CGS-2) dimer. The dose administered to the tumor by scFv (CGS-2) was also useful for scFv (CGS-2) ( 5 and 6 ) and correlated with their respective affinities (Table 1). However, both scFv (28SI) and scFv (CGS-2) appear prone to proteolytic cleavage and show high liver uptake (Fig. 6 ), while scFv (CGS-1) antibodies were significantly more stable and showed lower liver uptake ( 5 ).

LITERATURE QUOTES

• Ausubel et al. eds., John Wiley & Sons, 1992; Short Protocols in Molecular Biology, Second Edition.
• Alitalo et al. Adv. Cancer Res. 37, 111-158 (1982).
• Bagshawe K.D. et al. (1991) Antibody, Immunoconjugates and Radiopharmaceuticals 4: 915-922.
• Barone et al. EMBO J. 8, 1079-1085 (1989).
• Bird et al, Science, 242, 423-426, (1988).
• Borsi et al. J. Cell. Biol. 104, 595-600 (1987).
• Borsi et al. Anal. Biochem. 192, 372-379 (1991).
• Borsi et al. Int. J. Cancer 52, 688-692 (1992a).
• Borsi et al. Exp. Cell Res. 199, 98-105 (1992b).
• Brown et al. Amer. J. Pathol. 142, 793-801 (1993).
• Carnemolla et al. J. Cell Biol. 108, 1139-1148 (1989).
• Carnemolla et al. J. Biol. Chem. 24689-24692 (1992).
• Castellani et al. J. Cell. Biol. 103, 1671-1677 (1986).
• Castellani et al. Int. J. Cancer 59, 612-618 (1994).
• Crothers et al. Immunochemistry 9, 341-357 (1972).
• DeJager et al (1988) Proc. At the. Assoc. Cancer Res. 29: 377. Folli, et al (1994), Cancer Res., 54, 2643-2649.
• Ffrench-Constant et al. J. Cell Biol. 109, 903-914 (1989).
• Ffrench-Constant et al. Development 106, 375-388 (1989).
• Gibson TJ (1984) PhD thesis. (University of Cambridge, Cambridge, UK).
• Griffiths, et al. (1994), EMBO J. 13, 3245-3260. Gutman and Kornblihtt. Proc. Natl. Acad. Sci. (USA) 84, 7179-7182 (1987).
• Hawkins et al. Biol. 226, 889-896 (1992). Chubby et al Proc. Natl. Acad. Sci. USA 90 6444-6448, 1993).
• Holliger, P. and Winter G. Current Opinion Biotechnol. 4, 446-449 (1993).
• Hoogenboom et al. Nucl. Acids Res 19, 4133-4137 (1991).
• Humphries et al. J. Cell Biol. 103, 2637-2647 (1986).
• Huston et al, PNAS USA, 85, 5879-5883, (1988)
• Hynes Ann. Rev, Cell Biol. 1, 67-90 (1985).
• Jain RK. Sci. At the. 271, 58-65 (1994).
• Jonsson et al. BioTechniques 11, 620-627 (1991).
• Juweid et al. Cancer Res. 52, 5144-5153 (1992).
• Kaczmarek et al. Int. J. Cancer 58, 11-16 (1994).
• Kornblihtt et al. EMBO J. 4, 1755 (1985).
• Laitinen et al. Lab. Invest. 64, 375-388 (1991).
• Ledermann J.A. et al. (1991) Int J. Cancer 47: 659-664; low et al (1996), J. Mol. Biol., 260, 359-368.
• Marks et al (1991), J. Mol. Biol., 222, 581-597.
• Marks et al, (1992), Bio / Technology, 10, 779-783. McCafferty, J., Griffiths, A.D., Winter, G., Chiswell, D.J. (1990) Phage antibodies: filamentous phage displaying antibody variable domains. Nature (London), 348, 552-554.
• Neri et al (1996a), Bio / Techniques, 20, 708-713.
• Neri et al (1996b), Nature Biotechnology, 14, 385-390. Nissim et al. EMBO J. 13, 692-698 (1994).
• Norton and Hynes. Mol. Cell. Biol. 7, 4297-4307 (1987).
• Owens et al. Oxf. Surv. Eucaryot. Genes 3, 141-160 (1986).
• Oyama et al. J. Biol. Chem. 10331-10334 (1989).
• Oyama et al. Cancer Res. 50, 1075-1078 (1990). Pack et al (1993), Bio / Technology, 11, 1271-1277.
• Peters et al. Cell adhesions. Commun. 3, 67-89 (1995).
• Pierschbacher et al. Cell 26, 259-267 (1981).
• Pluckthun, A. Bio / Technology 9: 545-551 (1991).
• Reff, M.E. (1993) Curr. Opinion Biotech. 4: 573-576.
• Ruoslahti. Ann. Rev. Biochem. 57, 375-413 (1988).
• Saginati et al. Eur. J. Biochem. 205, 545-549 (1992).
• Schwarzbauer et al. EMBO J. 6, 2573-2580 (1987).
• Schroff et al., 1985 Cancer Res. 45: 879-885. Siri et al. Nucl. Acids Res. 19, 525-531 (1991).
• Tomlinson I.M. et al, (1992) J. Mol. Biol. 227: 776-798. Traunecker et al. Embo Journal, 10, 3655-3659, (1991).
• Trill J.J. et al. (1995) Curr. Opinion Biotech 6: 553-560. Vartio et al. J. Cell Science 88, 419-430 (1987).
• Vaughan et al (1996), Nature Biotechnol., 14, 309-314. Ward, IT. et al., Nature 341, 544-546 (1989).
• Winter, G and C. Milstein, Nature 349, 293-299, 1991.
• WO94 / 13804Yamada. Ann. Rev. Biochem. 52, 761-799 (1983).
• Zardi et al. EMBO J. 6, 2337-2342 (1987).

Claims (26)

An isolated antibody or isolated antibody fragment specific for and binding directly to the oncofetal ED-B domain of fibronectin (FN) and, when measured as purified monomer, has a dissociation constant (Kd) for ED-B-Fn of less than 6x 10 ^-8 M has. Isolated antibody or isolated antibody fragment according to claim 1, which is an antibody-antigen binding domain human Origin. Isolated antibody or isolated antibody fragment according to claim 1 which binds to any FN after treatment of the FN with the protease thermolysin ED-B. Isolated antibody or isolated antibody fragment according to claim 1, without the treatment of B-FN with N-glycanase binds to B-FN. An isolated antibody or isolated antibody fragment according to claim 1 which has a variable heavy chain (VH) region from that of DP47, codon 1 Glu up to and including codon 98 Arg in 1 which has human germline-derived sequence and the CDR3 sequence Ser Leu Pro Lys. An isolated antibody or isolated antibody fragment according to claim 1 which has a variable heavy chain (VH) region from that of DP47, codon 1 Glu up to and including codon 98 Arg in 1 which has human germline-derived sequence and the CDR3 sequence Gly Val Gly Ala Phe Arg Pro Tyr Arg Lys His Glu. An isolated antibody or antibody fragment as claimed in claim 1 which has a variable light chain (VL) region from that of DPL16, codon 1 Ser to codon 90 Ser in 1 which has human germ-line sequence and the remainder of the CDR3 sequence, Pro Val Val Leu Asn Gly Val Val. An isolated antibody or antibody fragment as claimed in claim 1 which has a variable light chain (VL) region from that of DPL16, codon 1 Ser to codon 90 Ser in 1 which has human germ-line sequence and the remainder of the CDR3 sequence, Pro Phe Glu His Asn Leu Val Val. Isolated antibody or isolated antibody fragment according to one of the claims 1 to 8, wherein the antibody or the antibody fragment a scFv molecule includes. Isolated antibody or isolated antibody fragment according to one of the claims 1 to 8, wherein the antibody or the antibody fragment comprises a dimeric scFv molecule. Isolated antibody fragment according to one of the claims 5 to 8. A pharmaceutical composition comprising a isolated antibodies or an isolated antibody fragment according to one of the claims 1 to 11 in an effective amount in association with a pharmaceutical acceptable vehicle. nucleic acid the for an isolated antibody or an isolated antibody fragment according to one of the claims 1 to 11 encoded. Phage who for an isolated antibody or an isolated antibody fragment according to one of the claims 1 to 11 encoded. A host cell containing nucleic acid according to claim 13 in vitro transformed or transfected. Isolated antibody or isolated antibody fragment according to one of the claims 1 to 11 for therapeutic use. Use of an isolated antibody or Antibody fragment according to one of the claims 1 to 11 in the manufacture of a medicament for targeting Tumors. Use of an isolated antibody or Antibody fragment according to one of the claims 1 to 11 in the preparation of a tumor imaging agent. Process for the production of an isolated antibody or Antibody fragment according to one of the claims 1 to 11, wherein the method according to the in vitro expression of nucleic acid Claim 13 in a host cell to the antibody or the antibody fragment to produce. A method of producing an isolated antibody or antibody fragment according to claim 1, said method comprising: a) screening a phage-expressed antibody fragment library with one of the fibronectin protein a recombinant fibronectin fragment containing the ED-B domain, wherein the antibody fragments are scFv molecules; b) infecting host bacterial cells with positive clones; c) subjecting positive phage clones to an affinity maturation process; d) repeating steps a) and b) to select positive phage clones with improved affinity for the antigen; e) infecting host cells with positive clones and purifying antibody molecules from these host cells. The method of claim 20 wherein the ScFv molecules are a Antibody-antigen binding domain human Origin. The method of claim 20 wherein step a) is the Screening with the recombinant antigens 7B89 or ED-B. A method according to any one of claims 20 to 22, wherein nucleic acid of a positive clones for production by in vitro expression of a product antibody or antibody fragment, for the the nucleic acid is used in a host cell, wherein the / the product antibody or Antibody fragment one Comprises antibody-antigen binding domain, the for the oncofetal ED-B domain of fibronectin (FN) and binds directly thereto. The method of claim 23, wherein the product antibody is as whole antibody is present. The method of claim 23, wherein the product antibody fragment is present as an scFv molecule. Diagnostic kit comprising an isolated antibody or an isolated antibody fragment according to one of the claims 1 to 11 and one or more reagents which determine the allow the element to bind to cells.